Efficient Suturing in Deformable Models

Abstract

Surgical procedures are among the most complex medical interventions routinely performed. The suturing task, ubiquitous to most surgical interventions, has a steep learning curve. With the increased focus on minimally invasive interventions, the suturing task has become more complex than ever and compulsory training on virtual simulators has become the norm. However, simulating the suturing task is not trivial, as it involves the simulation of needle-tissue and thread-tissue interaction, as well as contact between sutured boundaries. In this work, robust methods are proposed for e ciently simulating suturing using a linear elastic nite element-based approach. Important aspects of our proposed method include modeling contact conditions using complementarity relations and simulating needle driving in realtime, without introducing any new mesh elements, through the use of adjacency relationships inherent to manifold meshes and decomposing rigid body motion into sliding and deforming (sticking) components. In addition, a computationally e cient formulation of thread pulling is presented which relies on distance linearization and achieves plausible results. These techniques are implemented and tested in a cross-platform prototype system, which allows for interactive simulation of suturing with high-resolution models through a simpli ed anastomosis scenario while making use of multiple human-computer interaction devices and immersive rendering techniques.